Member log in

Website traffic: Nielsen drops monthly unique browser count

UPDATE Aug 20: Nielsen is retiring its monthly unique browser (UB) count from November 1.

The research company is encouraging publishers to switch to quoting daily or weekly figures before then (NBR has always quoted a weekly figure - which has recently topped 40,000, just quietly).

Nielsen points out, correctly, that it's cookie (browser-based tracking software) system leads to double-counting of a site's audience over time as people (or their security software or network admin) delete cookies, or they access the same website from different devices.

Nielsen recently supplemented its cookie-based tracking with a people panel (see below).

--------------------------------------------

With new website rating system, Nielsen aims to count humans, not browsers

July 17: Nielsen has launched a new "hybrid" online audience ratings system, which is says will be more accurate at counting actual humans.

The company has previous used tags (code on every website it measures) to report unique browsers and page impressions.

Now, that approach will be combined with a survey of the surfing habits of a panel of 3000 people, whose home connections are fixed with meters (so no porn now; the panel was recruited in 2011 and has gone through months of testing. The panel is weighted to be representative of the NZ online population as a whole, Nielsen says.)

The hybrid product is called, simply, Nielsen Online Ratings.

A rep for Nielsen said the price of the new service depended on several factors, including the size of a website publisher's existing business with the company.

Drawbacks of the old system
The "old" tag approach allows for a lot of granularity, such as detail on ratings for every story on every news site (at least for their publishers' eyes).

However, it is also relies on software "cookies" for tracking – and cookies can get deleted over time (either by a person manually or, more likely, by their browser, security software or a cookie cull initiated by their network admin).

Another drawback: the tag system counts individual browsers, as in web browsers. Visit a site from IE and Chrome, on the same PC, and you count as two browsers.

Double or triple counting occurs if the same person accesses a site from home and work, plus a smartphone or tablet.

Thus although NZ's population is around 4.4 million, if you tally all its stats for May (by Nielsen's count) you get 16 million unique browsers.

And similarly, while Nielsen reported 36,616 domestic unique browsers for the week of June 9 to 15 NBR ONLINE, our typical monthly figure is around 90,000 – nice, but not believable (the weekly figure is the one quoted in our media kit).

Drawbacks of a panel
A panel system has its own drawbacks.

This is not like television, where a big majority of Nielsen's panel can be relied on to watch the boob tube at home.

People surf the web at home, the office, school, varsity, internet cafe and, increasingly, on smartphones – and as Nielsen readily acknowledges, usage anywhere beyond home could get missed.

On the plus side, a panel is miles more accurate for measuring time spent on a website.

A hybrid appraoch 
With its new hybrid approach, Nielsen aims to meld panel and tagged code results to come up with a more realistic unique audience (UA) figure.

Nielsen will now report weekly figures from its "old" page tag system (Site Census and Market Intelligence) but will add new monthly figures based on data collected from the panel of 3000 people.

A bad cop would say it combines the weaknesses of both systems. It still uses cookies, and a panel just has the lingering smell of a home-user bias drifting over everything - even with some degree of weighting from tagged website results.

Initial results promising
Still, Nielsen's initial results are promising.

No publisher is adverse to puffed up results, of course, but when domestic unique browser results exceed NZ's population, it just gets silly.

Its first gob of monthly data says there is a unique 2+ audience of 3.3 million in NZ – that is, a figure that's actually within the bounds of possibility.

Trade Me - back down to earth
And Trade Me, which was given 6.1 million domestic unique browsers for May 2012 under the Nielsen Market Intelligence page tagging system, was reported as a far more realistic-sounding unique monthly audience of 1.88 million for the same month under the new hybrid system.

It's a step-down, but it gives Trade Me a monthly figure that its advertisers and investors can put more stock in.

There's also a new, TV-style audience reach figure (see the Top 10 chart above).

Similarly Stuff, which NBR understands is a Wellington-based site, registered 2.1 million domestic unique browsers for May 2012 under the Nielsen Market Intelligence page tagging system, is reported as having a unique monthly audience of 1.09 million under the hybrid system.

Any website publisher which has been quoting monthly rather than (more realistic) weekly figures from the old system will have to take their oats.

As well as the panel yielding far superior time spent on a site data (at least for sites accessed from home), the new system has frills like making it easier to filter out international traffic, and discounting for auto-refreshes.

Website publishers have been given a choice of sticking with the old system, or signing on for the new system (as a number already have).

Google drawn in
Unlike the tagging code, a panel is not opt-in. Any site its members are surfing too could feature in Nielsen's new monthly audience figures (publishers do have to subscribe to get fills like more detailed breakdowns for different parts of a site, trend information and so forth).

Nielsen has recommended to the IAB (Interactive Advertising Bureau - a group of major online publishers) that all of its members move to the new hybrid system.

Thus we get Google drawn into Nielsen's online survey for the first time (the search giant was a famous hold-out on the old system. Among other reasons, it was not about to add Nielsen code to its system, which slows things down factionally).

Don't look for NBR to feature at the top of the charts though.
 
Nielsen new monthly unique audience rankings (at least the public version) will sites that reach an (undisclosed) minimum sample size.
 
And as I've said before, there are only two stories in publishing – you've either got the biggest audience, or the best audience.
 
Not prizes for guessing which side of the equation your favourite paywall sits.
 

As a footnote: A number of IT and ad writers have been grumbling about the drawbacks of page tagging and a web browser-based system for years. Nielsen has openly acknowedged issues, but taken a while to address them. But if you go further back into the mists of time, a decade ago Nielsen operated its "@Home" panel of 3500 browsers. It went by-the-by after  Nielsen bought Red Sherriff and moved to centre is ratings around its new acquisition's Site Census browser approach.

 

 

More by this author

Comments and questions
3

Sounds as effective as peoplemeter data is for television viewership -

And a heads up for all the bloggers: all s13 sites on site meter have been down from about August 10. Rumours are the company is going down. So the bloggers rankings aren't going to mean much for a couple of months until we move to other counters Open Parachute can access.

Nielsen have just recently (earlier this week) announced that they are removing monthly UB figures from reporting data... Websites like us, that are still customers of the old system but reluctant to sign on to the Hybrid system are not being given the choice... The Hybrid system does not work for us, as most of our audience looks at our site during the weekday - not at night at home... GRRR...